Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/astral-sh/uv/load-github-action-threadnpx skills add astral-sh/uv --skill load-github-action-threadgit clone --depth 1 https://github.com/astral-sh/uvWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00062 | $0.00954 |
| Opus 5 | $0.00031 | $0.00477 |
| Sonnet 5 | $0.00012 | $0.00191 |
| Haiku 4.5 | $0.00006 | $0.00095 |
Grade A, and why
load-github-action-thread scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Load GitHub Action Thread
When given an issue instead of a run, find the related issue-triage run first. Resolve the issue title and number, then match the run display title:
repository=astral-sh/uv
issue=20477
issue_title="$(gh issue view "$issue" --repo "$repository" --json title --jq '.title')"
issue_number="$(gh issue view "$issue" --repo "$repository" --json number --jq '.number')"
gh run list --repo "$repository" --workflow issue-triage.yml --limit 500 \
--json databaseId,displayTitle,event,status,conclusion,createdAt,url \
| jq -c --arg title "$issue_title" '.[] | select(.displayTitle == $title)'
Use the newest matching run. An issue-triage run can contain both the triage and bug-reproduction threads. Also check for directly dispatched bug-reproduction runs; their display title does not include the issue title. If the issue title changed or a run was manually dispatched, list recent candidates and confirm them from their logs:
gh run list --repo "$repository" --workflow reproduce-bug.yml --limit 100 \
--json databaseId,event,status,conclusion,createdAt,url
gh run view <run-id> --repo "$repository" --log | rg -m 2 "ISSUE: $issue_number|issue: $issue_number"
When given a pull request, find its security-review run from the pull request checks. The check link contains the Actions run URL and can be passed directly to the loader:
pull_request=20482
gh pr checks "$pull_request" --repo "$repository" --json name,workflow,state,link \
--jq '.[] | select(.name | test("review"; "i"))'
The security-review job only runs for eligible same-repository pull requests; a skipped review has no thread artifact to load.
Run the repository utility with the GitHub Actions run URL:
./agents/scripts/load-github-action-thread.sh https://github.com/astral-sh/uv/actions/runs/123456789
The utility downloads every matching codex-thread-* artifact, finds the root codex exec rollout
for each step, and forks it into an interactive local Codex thread. Subagent rollouts are ignored;
they inherit the root prompt and would otherwise produce duplicate-looking tasks. The first line of
the root preview becomes the task title, including the step, source number, and issue or pull
request title. Forking is required because codex-action uses codex exec; copying an exec
rollout into ~/.codex/sessions alone does not make it visible in the Codex app.
For a numeric run ID, pass the repository when it cannot be inferred from the current checkout:
./agents/scripts/load-github-action-thread.sh 123456789 --repo astral-sh/uv
Use --artifact PATTERN to select a particular retained thread and --cwd PATH to attach the
imported history to a different local checkout:
./agents/scripts/load-github-action-thread.sh 123456789 \
--repo astral-sh/uv \
--artifact 'codex-thread-reproduce-bug-*' \
--cwd /path/to/uv
The utility requires gh, jq, and codex, plus zstd for compressed rollouts. It prints the
resolved title and a codex://threads/<id> link for every imported task and automatically opens the
Codex app when exactly one task is imported on macOS. For multiple steps, return a Markdown list
with a clickable link labeled by the corresponding task title without repeatedly switching the app:
- [Issue triage for #20477: Example issue](codex://threads/<id>)
- [Bug reproduction for #20477: Example issue](codex://threads/<id>)
Use the matching step title for codex-thread-pull-request-security-review-* and
codex-thread-release-prepare-*.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 89 lines · 62 tokens per session scan A 05f1f63e72e7
load-github-action-thread is a skill published in the GitHub repository astral-sh/uv (89,330 stars, last pushed today), licensed Apache-2.0. It adds 62 tokens to every session and 954 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
clawctl
Know the clawctl CLI and manage your Clawrium fleet (hosts, agents, providers, channels, integrations, skills, secrets).
dxt-packaging
MCPB (Model Context Protocol Bundle) Packaging and Distribution.
paperclip
Interact with the Paperclip control plane API for task coordination and governance. Use when checking assignments, updating issue status, posting comments, delegating work, managing routines, or calling Paperclip API endpoints.
omni-mcp
Connect to the OmniRoute MCP server (110 tools, 3 transports: SSE/stdio/HTTP). Covers routing, cache, compression, memory, skills, providers, and audit tools across 33 permission scopes.
add-resend
Add Resend (email) channel integration via Chat SDK.
codegen
Code generation utilities for json-render. Use when generating code from UI specs, building custom code exporters, traversing specs, or serializing props for @json-render/codegen.